Secure Multi-party Computation for Privacy-preserving Machine Learning in Healthcare

This study investigates the dual impact of Secure Multi-Party Computation (SMPC) on machine learning (ML) model performance and stakeholder trust within the context of healthcare data analytics. Using a structured survey, data were collected from 171 households in New York. The study employed R Studio for regression analysis and SPSS for descriptive statistics. Results reveal that SMPC significantly enhances ML model accuracy when combined with diverse datasets, indicating its effectiveness as a privacy-preserving solution. However, SMPC alone does not significantly increase stakeholder trust; rather, trust is strongly influenced by awareness of SMPC technology, perceived data privacy risks, and institutional reputation. These findings emphasize the importance of both technical and human-centric factors in adopting privacy-preserving analytics. The study offers valuable insights for healthcare organizations, policymakers, and technology developers seeking to balance privacy, performance, and trust in data-driven decision-making.